sol8 x86 dmamode help needed
J.D. Bronson
I have seen alot of threads on this and need some advice.
As you know, if you enable DMA on some motherboard/drive combinations a cold
boot will
work, but a warm reboot will hang.
Is there any total answer to this ? - I cannot change out the the
motherboard, but I did
notice something interesting:
IDE device at targ 0, lun 0 lastlun 0x0
model IBM-DTLA-307020, stat 50, err 0
cfg 0x45a, cyl 16383, hd 16, sec/trk 63
mult1 0x8010, mult2 0x110, dwcap 0x0, cap 0x2f00
piomode 0x200, dmamode 0x200, advpiomode 0x3
minpio 240, minpioflow 120
valid 0x7, dwdma 0x7, majver 0x3c
and
IDE device at targ 0, lun 0 lastlun 0x0
model Maxtor 5T020H2, stat 50, err 0
cfg 0x40, cyl 16383, hd 16, sec/trk 63
mult1 0x8010, mult2 0x110, dwcap 0x0, cap 0x2f00
piomode 0x200, dmamode 0x0, advpiomode 0x3
minpio 120, minpioflow 120
valid 0x7, dwdma 0x7, majver 0x7e
'dmamode' is the key here. The first drive shows '0x200' and the second
shows '0x0' . The second drive will reboot
just fine.....
So, if anyone can recommend a really good fix I would appreciate it. SCSI is
not an option.
Or, if someone knows a recent ~40GB drive that gives out a '0x0' I would
consider that as a work around.
The maxtor drive shown above is older and only 20GB.
Please help?
JBronson
{jb}[at]-[ktxg]DOT{com}
Juergen Keil
> As you know, if you enable DMA on some motherboard/drive combinations a cold
> boot will
> work, but a warm reboot will hang.
>
> Is there any total answer to this ? ...
>
> IDE device at targ 0, lun 0 lastlun 0x0
> model IBM-DTLA-307020, stat 50, err 0
See
http://groups.google.de/groups?selm=wyitm6nxj9.fsf%40leo.tools.intra
Bruce Adler's patch works for me with all sorts of IBM IDE drives.
J.D. Bronson
"Juergen Keil" <j...@tools.de> wrote in message
news:wy1y7je...@tools.de...
Which is correct?
ata_devo_reset+0x54?w 0xaa68 ;; was 0xcc68
ata_devo_reset+0x54?w 0x6668 ;; was 0xcc68
The patch Bruce posted or the one the next poster posted?
Juergen Keil
Well, in the same thread Bruce explained that both versions work.
I'm using the 0x6668 one.
John D Groenveld
> http://groups.google.de/groups?selm=wyitm6nxj9.fsf%40leo.tools.intra
>
>Bruce Adler's patch works for me with all sorts of IBM IDE drives.
Bruce's patch works in my Toshiba Tecra 8100 with a IBM-DARA-212000 disk.
John
groe...@acm.org
John D Groenveld
>Well, in the same thread Bruce explained that both versions work.
>I'm using the 0x6668 one.
The 0x6668 did not work for me, but 0xaa68 did.
BugID: 4429794, "ATA: sys w/ AWARD BIOS, IBM ultra-DMA disk hangs
after soft reboot" advocates the 0x6668 patch.
John
groe...@acm.org
J.D. Bronson
"John D Groenveld" <groe...@cse.psu.edu> wrote in message
news:amq6tj$m0j$1...@ewok.cse.psu.edu...
I now have a new issue that neither 0x6668 or 0xaa68 addresses and I dont
know what it means:
Sep 25 16:44:34 xeon unix: IDE device at targ 0, lun 0 lastlun 0x0
Sep 25 16:44:34 xeon unix: model Maxtor 5T020H2, stat 50, err 0
Sep 25 16:44:34 xeon unix: cfg 0x40, cyl 16383, hd 16, sec/trk
63
Sep 25 16:44:34 xeon unix: mult1 0x8010, mult2 0x110, dwcap
0x0, cap 0x2f00
Sep 25 16:44:34 xeon unix: piomode 0x200, dmamode 0x0,
advpiomode 0x3
Sep 25 16:44:34 xeon unix: minpio 120, minpioflow 120
Sep 25 16:44:34 xeon unix: valid 0x7, dwdma 0x7, majver 0x7e
Sep 25 16:44:36 xeon unix: PCI-device: ide@0, ata0
Sep 25 16:44:36 xeon unix: ata0 is /pci@0,0/pci-ide@7,1/ide@0
Sep 25 16:44:36 xeon unix: IDE device at targ 0, lun 0 lastlun 0x0
Sep 25 16:44:36 xeon unix: model ST330011A, stat 50, err 0
Sep 25 16:44:36 xeon unix: cfg 0xc5a, cyl 16383, hd 16, sec/trk
63
Sep 25 16:44:36 xeon unix: mult1 0x8010, mult2 0x110, dwcap
0x0, cap 0x2f00
Sep 25 16:44:36 xeon unix: piomode 0x200, dmamode 0x200,
advpiomode 0x3
Sep 25 16:44:36 xeon unix: minpio 240, minpioflow 120
Sep 25 16:44:36 xeon unix: valid 0x7, dwdma 0x7, majver 0x3e
Sep 25 16:44:36 xeon unix: ata_set_feature: (0x66,0x0) failed
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
I only see this on the 2nd Drive (Seagate Barracuda IV)
Help?!
Bruce Adler
"J.D. Bronson" <nos...@for.me> wrote in message news:up4bpbn...@corp.supernews.com...
> ...
> Sep 25 16:44:36 xeon unix: ata_set_feature: (0x66,0x0) failed
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> I only see this on the 2nd Drive (Seagate Barracuda IV)
> Help?!
Just ignore it. Given that you're not seeing any sort of problem or
resets the lack of that drive feature is irrelevant.
The fact that your drive doesn't support that feature is only of interest
if your drive fails so badly that the driver decides a software reset is
necessary. Drives that don't support that feature almost never recover
correctly after a software reset. Drives that do support that feature are
*sometimes* (but not always) able to recover after a software reset.
Knowing which kind of drive you have would help Sun to respond to any
bug report you file about error recovery not functioning correctly.
J.D. Bronson
"Bruce Adler" <bruce.NxOxSx...@acm.org> wrote in message
news:6Hqk9.3198$u56.2...@newsread1.prod.itd.earthlink.net...
Thanks Bruce. After all, without your patch, we would not be able to run
Solaris :)